摘要
纹理在自然图像中普遍存在的,纹理分析一直是计算机视觉领域中的一个重要研究方向。在企业中铝型材表面喷涂质量的检测都是通过人眼检测的。从计算机视觉的角度出发,提出基于纹理分析的铝型材表面喷涂质量检测方法。首先采用了灰度共生矩阵和Gabor滤波分别进行提取纹理特征,然后通过纹理特征分析区别出大砂和细砂产品,从而达到质量检测的效果,最后将灰度共生矩阵、Gabor和神经网络的分类精度进行对比,发现基于Gabor滤波的纹理分析方法对铝型材喷涂表面图像分类效果明显。
Texture is widespread in natural images.Texture analysis has been an important research direction in the field of computer vision. Aluminum surface painting quality is inspected by the human eye in the enterprise. From the perspective of computer vision,this paper puts forward the method of aluminum surface coating quality inspection based on texture analysis. The gray level co-occurrence matrix and Gabor filter are used to extract feature. The texture feature difference between sand and fine sand products is used to achieve the quality inspection. Finally the gray level co-occurrence matrix, Gabor and neural network classification precision are compared. It is found that the texture analysis method based on Gabor filter can obtain obvious image classification effect of the surface on aluminum coating.
出处
《机械工程师》
2015年第10期80-82,共3页
Mechanical Engineer
关键词
纹理分析
灰度共生矩阵
GABOR变换
质量检测
texture analysis
gray-level co-occurrence matrix
Gabor transform
quality inspection
